Cluster Match QF Announcement

GM all public goods funders!

I’ve got some exciting news to share regarding quadratic funding at Giveth. We have been experimenting with novel QF-related mechanisms and have decided to utilize a variation of QF called Connection-Oriented Cluster Match (COCM) in our rounds going forward, starting with the Galactic Giving round.

COCM offers a more systematic and democratic approach to distributing matching funds for QF, potentially improving Sybil defense while more effectively rewarding cooperation across social differences.

Why Improve QF?

Giveth has operated 6 quadratic funding rounds to date and overall they have been a resounding success. QF has proven its validity as a useful mechanism for distributing a pool of funds to projects based on the preferences of donors and rallying more support for impact projects working hard to make a difference.

However, QF in its current form is far from a perfect fit for the world we live in. In order for QF to work perfectly, we need to assume everyone is a 100% rational actor and acts solely for their own benefit, i.e. an “atomized, rational agent” in economics terms. We know this will never be the case, as humans are complex social creatures.

As a result, regular QF is susceptible to coordination imbalances and collusion, but there is active research going on into how QF-related mechanisms can be adapted to account for the influence of social connections while incentivizing cooperation across social differences.

One of the most promising new QF mechanisms is called Connection-Oriented Cluster Match, which Giveth has begun experimenting with.

Connection-Oriented Cluster Match QF

Connection-Oriented Cluster Match (COCM) is a relatively new mechanism, first proposed in December 2022 in a research paper titled Beyond Collusion Resistance: Leveraging Social Information for Plural Funding and Voting.

Gitcoin has successfully implemented this mechanism into its quadratic funding rounds, and the results appear promising. Giveth aims to build upon this success by implementing COCM into our QF round workflow as we believe it will yield better results for projects, donors, and Giveth contributors alike. We will also continuously test and tweak the algorithm to ensure it works best for the way Giveth runs QF rounds and the goals we want to achieve.

So, what is it? COCM is a mechanism used to identify groups or clusters of entities within a network based on their similarities or connections. In the context of Giveth QF rounds, COCM will involve analyzing the donation behavior between users participating in funding rounds to detect potential Sybil attacks and allocate matching funds in a way that rewards unique cooperation. By grouping users with similar donation patterns, COCM helps identify legitimate contributors and mitigate the influence of Sybil attacks.

Preliminary Results

The Galactic Giving round (live May 2-16, 2024) will be the first time that Giveth uses COCM to distribute matching funds but we have been performing preliminary experiments using the data from our previous rounds:
QF Round Matching (Regular QF)

Y axis - matching funds
X axis - projects (A through HH)

QF Round Matching (COCM) (2)

As you can see in the comparison of these two charts, COCM, when compared to our regular implementation of QF, “took away” some of the matching funds going to top projects and redistributed them across other projects. The top projects in this case likely have donors with very similar donation patterns (e.g. they chose to support the same limited number of projects). When using COCM, these donors are clustered together and their donations count as one donation from one unique donor, so the result is a lower impact on the matching fund distribution.

With COCM, the funds are distributed more evenly across a wider array of projects. We also see a longer tail to the right, meaning fewer projects receive $0 or negligible matching.


Giveth continues to evolve and position itself as an industry leader in the quadratic funding space. This new COCM mechanism experimentation is a big step forward and we are confident it will result in a better QF round product for everyone involved. The Galactic Giving round will be the first time we implement COCM (barring any serious issues with testing, in which case we’d revert to regular QF).

Please feel free to add questions/comments/suggestions below, and be on the lookout for more updates on QF experiments going forward!

I want to thank Glen Weyl and Joel Miller for their research on QF and related pluralistic mechanisms that gave life to this concept. Also thank you to Umar Khan who is working on implementing and improving cluster matching at Gitcoin and developing great open-source tools. To learn more check out the GreenPill Cluster Mapping Episode w/ Owocki, Joel, and Umar.


Thanks so much @koday for getting this up and for leading the charge for the implementation of COCM. The distribution you shared using the previous round’s data, I think the final results are looking like really really promising, and I think that cluster mapping is going to make a huge impact on improving our ability to distribute matching funds in a more democratic and fair way… and I’m excited to see a more diverse set of projects meeting their funding goals!


So excited for this, it is a HUGE thing, it reduces our costs to run the round by A LOT! Not just the 5% of the round size since we don’t need to out source this work to TrustaLabs, but also because the work to analyze TL’s data for sybils often costs us 10-60 hours from Lauren, Keiran and I…

This upgrade might actually make running these rounds PROFITABLE for the DAO, which, in previous rounds, it really wasn’t.

And then on top of that… the distributions, (from a subjective POV when you look at the real project names), just look more accurate!! The obscure projects that rally their community HARD do a lot worse, and the projects that are just generally appreciated, get much higher results. It’s beautiful.



Hi all, this is Swift Evo from DeSci Asia (we are participating in GIVETH QF now). I worte a chinese article about what is QF and cluster match.

Feel free to share it to Chinese GIVER in GIVEH.

1 Like

This is awesome! Thanks for writing this article about QF and Cluster Match and thanks for sharing it here :slightly_smiling_face: